eHam Forums => Boat Anchors => Topic started by: WB4SPT on April 06, 2012, 03:27:30 PM

Title: Why triple conversion?
Post by: WB4SPT on April 06, 2012, 03:27:30 PM
  I've admired the Hammarlund receivers for decades now, never owned one.  I did pick up a Drake R4A recently.  So, here is the leading question.  What does the triple conversion scheme do for the Hammarlund?  Does it have better IM or image than the ten year newer Drake?  The Drake is speced at better than 60dB.  Even more to the point:  Will one of the better Hammarlund's receive better under some conditions than the Drake?  Obviously, the Drake is not out of the box continuous receive, and would take a bunch of crystals to make it so.  

Title: RE: Why triple conversion?
Post by: W9GB on April 06, 2012, 03:56:44 PM
David -

Some of this is covered in earlier ARRL Handbooks, that taught you receiver design principles.
Edwin Armstrong's 1918 Superheterodyne receiver patents & design is the baseline for your question.

Introduction to the Superheterodyne Receiver
by Lloyd Butler, VK5BR

As these general coverage receivers continued to increase in frequency,
first to 10 MHz, then 30 MHz, and then VHF up to 50 MHz ....
The receiver IF frequencies also increased leading to multiple conversion schemes.

Ironically, today --- you see Direct Conversion & Single Conversion receiver designs providing
improved performance over the single conversion designs from decades earlier.

Title: RE: Why triple conversion?
Post by: WB6BYU on April 06, 2012, 07:16:02 PM
When your selectivity is determined by LC circuits you need a low IF to get a narrow
bandwidth.  Frequencies below 100kHz were often used for this reason.  (The bandwidth
is some percentage of the frequency for a given Q.)

It's difficult to get good image rejection when the first IF is less than about 1/10 of
the upper frequency of the receiver.  So if you are tuning up to 30 or 60MHz, you
end up with the first IF in the 4 - 12 MHz range.

So you end up with double (or sometimes triple) conversion to satisfy both requirements.
Sometimes you need an extra stage of conversion for things like keeping your BFO signal
out of your AGC loop.

Using crystal filters it is possible to get similar selectivity at higher frequencies.  So a single
IF at 5 to 9 MHz can give you sufficient selectivity, and a single conversion receiver
becomes more practical.  And when you are designing a receiver to transceive with
a matching exciter, that allows you to use the same oscillators and filters in both.

It isn't the number of conversions that makes a receiver good or not - it is how well
they are implemented.  Multiple conversions require more work to eliminate birdies
and close attention to signal levels to avoid overload in the later mixers.  You can
build good or awful receivers using either approach.  The Hammurlund receiver designs
are older and weren't originally designed to use crystal filters (though some could be
fitted with mechanical filters at 455 kHz or so.)  The Drake R4A was designed to operate
SSB along with the matching transmitter using crystal filters.

Different approaches to a similar result from different eras.

Which is better?  Can't say, especially at their current age.  Depends a lot on their
individual specifications and which ones are most important for your operating.
I know there are a number of mods to improve the Drake receivers.

Title: RE: Why triple conversion?
Post by: G3RZP on April 07, 2012, 04:29:34 AM
The HQ170 and 180 used a 3035 or 455 kHz first IF, depending on the frequency it was tuned to. The third IF was 60kHz, and it would be hard to get acceptable image rejection with this when down converting from 3035kHz.

Generally, multiple conversions were done to get better selectivity and image rejection. But the problem was signal handling, as there was generally too much gain before the selectivity.

Title: RE: Why triple conversion?
Post by: K1ZJH on April 07, 2012, 06:25:50 AM
I'm surprised more companies didn't use phasing to eliminate the image frequency when going between IFs.
I believe the Drake R8 was a pioneer in that regard.


Title: RE: Why triple conversion?
Post by: G3RZP on April 08, 2012, 03:53:07 AM
It's very hard to get consistency and not too much drift with time and temperature. Even when you are doing it on an integrated circuit with various feedback loops built in, a consistent 40dB takes some getting.

Incidentally, I've lost the reference but the idea of the superhet was produced in France in 1909, but the technology didn't exist to implement it. Same as the FET - invented in 1928, but impossible to manufacure at that time.

Title: RE: Why triple conversion?
Post by: AC5UP on April 08, 2012, 06:37:52 AM
This reminds me of a newsletter article I wrote a few years back which was intended to tickle the grey matter with the question: " How many different electronic parts are there? "

My answer was surprisingly few, as the difference between a generator, alternator, synchronous AC motor, transformer, choke and solenoid is mostly in how the windings of enameled copper wire are arranged. If a transistor is essentially two PN diodes arranged nose-to-nose, is an IC like an incredibly large diode array? You get the idea..........

A superheterodyne receiver is often thought of as a wholly unique design. I say not so much, and here's why: Look at the earliest designs like the TRF and Neutrodyne. Very simple concept with 1, 2, 3 or 4 variable tuned LC networks at RF with a triode between each stage for amplification. Each additional stage added cost, gain and selectivity along with another knob that needed adjustment whenever the station was changed. The number of knobs and switches made early radios intimidating to the average consumer but were like a playground for the geeks of the day.

Then the superheterodyne came along. Two knobs. Tuning and on/off volume. What could be simpler? And only a radical design change could take a radio from five knobs to two... Right?

Except, when you think about it, a superheterodyne is almost exactly like a TRF south of the converter stage. The IF and detector stages comprise a fixed-tune receiver often parked at 455 KC's and the audio stages are essentially the same. So what's different? The converter stage translates the RF frequency down to the fixed IF frequency. In a TRF design the entire radio is tuned to the RF frequency. Instead of everybody moving when the station is changed, now only one stage has to QSY.

In my mind that's not exactly a major design revision, and what was the technology that made it possible? The idea of adding a few extra grids to a triode. If one of the grids is provoked into oscillation, a pentagrid converter tube will mix that signal with any RF on another grid...

And should the product of the two come out at 455 KC's, there will be radio.          ;)

BTW: I rarely use the spell checker on this board but tried it today. The word " triode " is not in the eHam dictionary. Tsk!

Title: RE: Why triple conversion?
Post by: K1ZJH on April 08, 2012, 07:18:57 AM
The earliest superhets were all triodes, and battery powered.  199s, '01As.

The biggest obstacle was RCA held the patents.  Many early superhets were made from
kits to get around RCA's hold.


Title: RE: Why triple conversion?
Post by: G3RZP on April 08, 2012, 08:20:47 AM
I have never understood how the RCA stranglehold on so many tube patents didn't run them into trouble under the anti-trust laws. Can any of you guys explain?

Title: RE: Why triple conversion?
Post by: AC5UP on April 08, 2012, 10:36:44 AM
The earliest superhets were all triodes, and battery powered.  199s, '01As.

You're right... I'm so used to seeing a 6SA7 or 6BE6 pentagrid converter in a squidmatic I forget that it can be done with triodes. Plenty of early transistor radios did the same with PNP Germanium parts and that's the functional equivalent of a triode.

As for why RCA could hold so many patents, the only absolute law in the US of A is the greenback dollar and RCA held plenty of those as well. They were like the Microsoft of the 30's and 40's. Look up the history of the company, but also pay attention to how "General" David Sarnoff got that rank. You should also note the difference between the RCA Red and RCA Blue radio networks was something only a lawyer could fully appreciate and the way the work of Edwin Armstrong and Philo T Farnsworth was respected by RCA.

The old saying is "money talks, bullshiz walks" and it was only a few years ago when the Attorney General of the United States testified under oath that he really couldn't say if waterboarding constituted torture. Just wasn't sure, and that was OK. Oh, and we also have a supreme court that believes corporations are people and people can be strip searched for any reason while in custody by law enforcement. Personally, I'm waiting to see a corporation get strip searched while under investigation for a contract breach or similar since a corporation is a person, but I have a hunch it won't happen in my lifetime..........    ;)

Title: RE: Why triple conversion?
Post by: G3RZP on April 09, 2012, 03:00:01 AM

You are nearly as big a cynic as I am!

Title: RE: Why triple conversion?
Post by: AC5UP on April 09, 2012, 04:27:15 PM
...and with good reason, I think:

Look for the purple speaker icon toward the upper left to listen.

Title: RE: Why triple conversion?
Post by: N2EY on April 09, 2012, 06:00:42 PM
...and with good reason, I think:

Remember the wisdom of Solomon in the book of Proverbs. "He that troubleth his own house shall inherit the wind."

73 de Jim, N2EY

Title: RE: Why triple conversion?
Post by: KM1H on April 14, 2012, 12:43:00 PM
The more conversions the more noise and overload problems.

There are few ham level receivers that implemented a very low 2nd IF double conversion well much less triple. National seems to have excellent image specs with their 2215/80 kc in the NC-300/303 and my on the air results have been very positive. The SX-115 triple conversion is far superior to the Hammarlund in performance and ease of alignment IMO.